lable at ScienceDirect

Journal of Cleaner Production 278 (2021) 123771

Contents lists avai

Journal of Cleaner Production

journal homepage: www.elsevier .com/locate/ jc lepro

Rice-pasture agroecosystem intensification affects energy useefficiency

Ignacio Macedo a, d, *, Jos�e A. Terra a, Guillermo Siri-Prieto b, Jos�e Ignacio Velazco c, d,Leonidas Carrasco-Letelier d, **

a Instituto Nacional de Investigaci�on Agropecuaria (INIA), Programa Nacional de Arroz, Estaci�on Experimental INIA Treinta y Tres, Ruta 8 Km 282, CP 33000,Treinta y Tres, Uruguayb Estaci�on Experimental Mario Cassinoni (EEMAC), Facultad de Agronomía, Universidad de La República, Paysandú, Uruguayc Instituto Nacional de Investigaci�on Agropecuaria (INIA), Programa Nacional de Carne y Lana, Estaci�on Experimental INIA Treinta y Tres, Ruta 8 Km 282, CP33000, Treinta y Tres, Uruguayd Instituto Nacional de Investigaci�on Agropecuaria (INIA), Programa de Producci�on y Sustentabilidad Ambiental, Estaci�on Experimental INIA La EstanzuelaAlberto Boerger, Ruta 50 Km 11, 70000, Colonia, Uruguay

a r t i c l e i n f o

Article history:Received 11 October 2019Received in revised form23 July 2020Accepted 15 August 2020Available online 29 August 2020

Handling editor: Prof. Jiri Jaromir Kleme�s

Keywords:Life cycle assessmentCrop rotationCover cropsPerennial pastureIntegrated agroecosystems

* Corresponding author. Instituto Nacional de InvesPrograma Nacional de Arroz, Estaci�on Experimental IN282, CP 33000, Treinta y Tres, Uruguay.** Corresponding author.

E-mail addresses: imacedo@inia.org.uy, macedoyalcarrasco@inia.org.uy (L. Carrasco-Letelier).

https://doi.org/10.1016/j.jclepro.2020.1237710959-6526/© 2020 Elsevier Ltd. All rights reserved.

a b s t r a c t

Sustainable rice production systems are key to food security. Diversified farming systems are essential forecological intensification and environmental enhancement. Energy use efficiency is one of the mainsustainability indicators in agroecosystems. Thus, an assessment of consumption and efficiency of energyin contrasting cropping systems can discriminate their management practices and components sus-tainability. The goal of this study was to evaluate the energy performance through energy return oninvestment (EROI) in four rice-based rotation systems that belong to a long-term experiment located inthe Temperate Grassland Terrestrial Ecoregion, at the Atlantic side of South America. Rotations analyzedconsisted in: a) continuous rice (Rc); b) rice-soybean (R� S); c) rice-pasture for 1.5 years (R� PS); and, d)rice-pasture for 3.5 years (R� PL). The EROI estimations considered all the inputs and outputs of energyfrom cradle to farm gate. The greatest EROI was observed in ReS (7.2 MJ MJ�1) and the lowest energyconsumption in R� PL (10,607MJ ðha yrÞ�1). The R� PL’s EROI (6.7MJ MJ�1) was 6.5% and 8% higher thanRc and R� PS EROI, respectively. Rotations without pastures produced 79% more energy compared withrotations including pastures. However, energy inputs of rice-pasture rotations were 40% lower thaneither R� S or Rc. The EROI (without animal production) of R� PS , ReS and Rc was 25%, 28% and 43%lower than the EROI of R� PL (10 MJ MJ�1), respectively. For the analyzed South American ecoregion,EROI assessments of four business as usual rice production systems allowed to discriminate and hier-archize their sustainability and diversity.

© 2020 Elsevier Ltd. All rights reserved.

1. Introduction

Global population growth and its food demands (Godfray et al.,2010; Tilman et al., 2011) are pushing for a 90% increase inworldwide agricultural production by 2050 (FAO, 2009). This haschallenged the agricultural sector to find sustainable technological

tigaci�on Agropecuaria (INIA),IA Treinta y Tres, Ruta 8 km

por@gmail.com (I. Macedo),

options for expansion and intensification, amid a climate changescenario. Rice is a major crop worldwide, mostly cultivated incontinuous cropping systems in Asia (Seck et al., 2012) but also inother more diverse systems in the world (Bryant et al., 2012;Martins et al., 2016), being strategic for global food security (Secket al., 2012). Sustainable increase of rice production implies theredesign of agricultural systems to mitigate environmental impactsand increase efficiency and ecosystems services (FAO, 2009).Ecological intensification examples of rice production sharing oralternating the land use are the rice-fish co-culture systemsdeveloped at asian regions (Conway, 2012; Wan et al., 2019b), therice-soybeans systems in Arkansas USA and Brazil (Bryant et al.,2012; Martins et al., 2016), and the rice-beef rotation systems

Fig. 1. Political map of Uruguay: rice cropping land use (solid pink polygons) and theLTE-RC location (black triangle). (For interpretation of the references to colour in thisfigure legend, the reader is referred to the Web version of this article.)

I. Macedo et al. / Journal of Cleaner Production 278 (2021) 1237712

developed in Temperate Grassland terrestrial ecoregion of SouthAmerica)(Bao et al., 2018; Olson et al., 2001; Pittelkow et al., 2016).These production systems could be alternatives to develop andachieve ecological intensification of agroecosystems.

Intensification of any process requires an increase in energydemands (Maraseni et al., 2015). In the case of agricultural inten-sification, any increase of energy efficiency should increase globalenergy consumption; situation that is known as the Jevons paradox(Pellegrini and Fern�andez, 2018). In this sense, not always anintensification of agriculture production could be environmentallypositive. Today in agriculture, energy consumption is higher than itwas before the Green Revolution, due to irrigation, fertilizers,pesticides, fuels, machineries, etc. (Conforti and Giampietro, 1997;Woods et al., 2010). Energy consumption among some of Africa’sagricultural systems is less than 1000 MJ (MJ) per hectare (ha)compared with values of 30,000 MJ ha�1 observed in the UnitedStates (Pimentel, 2009).

It is well known that current intensification models have manynegative externalities (Godfray et al., 2010; Tilman et al., 2011). Inthe future, not only will be necessary to meet food demands, butregulation ecosystem services such as soil erosion control, climateregulation, and pest regulation will also require consideration. Theconcept of ecological intensification has been promoted in order toredesign agroecosystems using landscape approaches and consid-ering the natural functionalities of ecosystems (Tittonell, 2014).Thus, differentiating systems for environmental sustainability is achallenge when designing new agricultural systems.

Two main factors improve energy performance and agricultureyield: reducing energy inputs (EI) and/or increasing energy equiv-alent produced (EP), the latter being embodied in the harvestedbiomass (Swanton et al., 1996). Therefore, the goal is to improve theenergy return on investment (EROI), or the ratio between EP and EI.

The EROI performance assessment has focused primarily onsingle crops (Franzluebbers and Francis, 1995; Pittelkow et al.,2016). Few studies have evaluated either the effects of manage-ment practices (i.e. crop rotations) (Rathke et al., 2007) or theirenergy performance (i.e. production systems, rotations) (Theisenet al., 2017). Crop rotations were designed to minimize risks,maintain soil fertility, and interrupt weeds, pests, and diseasescycles (Bird et al., 1990; Liebman and Janke,1990). Furthermore, theinclusion of legumes could reduce nitrogen (N) requirements fromcommercial fertilizers (Heichel and Barnes, 1984).

There are many models of ecological intensification, usingvarious technological options, such as border crops in urban agri-culture or rice-fish coculture systems (Wan et al., 2018, 2019b);among their other effects, these system types are able to reducepesticide and nitrogen fertilizer use. Diversified farming systemsand some forms of conservation agriculture represent other modelsof ecological intensification (Tittonell, 2014).

Integrated crop-pasture rotation systems are rare globally,except in some regions of Argentina (Díaz-Zorita et al., 2002),southern Brazil Carvalho et al. (2018); de Moraes et al. (2014), andUruguay (García-Pr�echac et al., 2004). Compared with continuousrice systems, rice-pasture rotation systems allow farmers to sustainhigh productivity, maintain soil quality, diversify incomes, andminimize the use of fertilizers and pesticides (Deambrosi, 2003;Pittelkow et al., 2016). For example, a rice crop grown after aperennial grass-legume mix pastures in Uruguay is fertilized with60e80 kg N ha�1, 40e50 kg P2O5 ha�1, and 30e40 kg K2O; fungi-cides are generally used once in the crop cycle and insecticides arevirtually disused. Symbiotic nitrogen fixation of pasture legumesminimizes synthetic nitrogen fertilizer use and has indirect impactson energy consumption (Heichel and Barnes, 1984). However,global market demands are pushing to reduced pasture areas andincrease other crops, such as soybeans, within rotation systems.

Thus, following short-term economic profits, the Uruguayan ricesystem has developed different technological options to reduce theproportion of perennial pastures within rotations (Deambrosi,2003; Pittelkow et al., 2016). However, it is unclear whetherthese intensification processes are sustainable.

A long-term rice production system experiment (LTE-RC) wasinstalled at the Uruguayan National Institute of AgricultureResearch (INIA) to evaluate the environmental sustainability ofdifferent rotations. The hypothesis of this study was that theenvironmental sustainability performance of long-term rice rota-tion systems can be assessed using EROI, to rank ecological inten-sification pathways. The EROI of four rice-based cropping systemswith different intensification levels was estimated for: rice inrotation with long-term pastures, with short-term pastures, withsoybean crop, and continuous rice.

2. Materials and methods

In 2012, the LTE-RC was installed at the INIA Paso de la LagunaExperimental Station in Treinta y Tres, Uruguay (33+602300 S,54+1002400 W; located 22 m above sea level) (Fig. 1). Mean annualrainfall at the site is 1360 ± 315 mm; annual total potentialevapotranspiration is 1138 ± 177 mm; mean monthly temperatureis 22.3 ± 0.85+C and 11.5 ± 0.82+C during summer and winter,respectively. The dominant soil at the site is an Argialboll with aslope less than 0.5% (Dur�an et al., 2006).

2.1. Rice cropping systems

The LTE-RC evaluated six rice cropping system (i.e. rotation)intensities under no-till conditions. Cropping intensity treatmentswere differentiated based on the proportion and length of pasturesvs. rice, and on crop phases in the rotation. For this study, four of the

Table 2Profile of nutrients used in each crop rotation system. Values are expressed askg ðha yrÞ�1.

Treatment Crop N P2O5 K2O

RC rice 166 75 51R� S rice 79 61 63

soybean 5 91 45R� PS rice 99 15 72

pasture 23 45 22R� PL 1st rice 79 15 31

2nd rice 83 15 44pasture 8 46 8

I. Macedo et al. / Journal of Cleaner Production 278 (2021) 123771 3

six contrasting rotations were selected (Table 1).Treatments were established in 20-mwide and 60-m long plots

in a randomized complete block design with three replications; allrotation phases were present each year (Patterson, 1964). Crop andpasture management, including fertilization and pesticide appli-cation, followed INIA Rice Program recommendations. All crop andpasture operations (including seeding, fertilization, pesticideapplication, and harvesting) at each experimental unit weremanaged with machinery similar to that used by farmers. Routinecrop management and performance data were recorded by INIA’sRice Program team. The crop rotations described above wereevaluated in seasons 2015e2016 and 2016e2017. Rice was seededin mid-October and soybean at the beginning of November. Trifo-lium alexandrinum and Lolium multiflorum (ryegrass) were seededas cover crops immediately following rice and soybean harvests inApril.

2.2. Rice cropping fertilization and pesticide use

Phosphorus (P) and potassium (K) fertilization in continuousrice (RC) and rice-soybean (ReS) was equivalent to the previouscrop extraction, while rice crops in rotation with pastures for 1.5years (R� PS) or 3.5 years (R� PL) were fertilized according tosufficiency levels. Critical levels were 7 mg of P kg�1 of soil(Hern�andez et al., 2013), and 78.2 mg of K kg�1 of soil at pH 7(Deambrosi et al., 2015). Cover crops T. alexandrinum andL. multiflorumwere not fertilized. Fertilization with nitrogen in theRc system was applied as needed to reach a yield of 10 t ha�1. Inother rotations, rice was fertilized with nitrogen based on criticalsoil nitrogen mineralization potential levels developed by Castilloet al. (2015). In all cases, nitrogen was split into two applicationsas urea: at tillering (dry application) and immediately beforepanicle initiation (flooded soil). Doses of N, P2O5, and K2O used foreach rotation and crop are described in Table 2.

Fertilizer and pesticide management was based on INIA RiceProgram recommendations, as shown in Table 3.

2.3. Irrigation

Irrigation was applied using a 150 HP diesel pump with a flowrate of 220 L h�1 and diesel consumption of 14 L h�1. Water con-sumption for rice cultivation was assumed in this experiment to be10,000 m3 ha�1 (Carracelas et al., 2019). Therefore, the estimatedconsumption of diesel oil by irrigation was 176.76 L ha�1.

Table 1Crop rotation systems, annual and seasonal schedules (SS, spring-summer; AW, autumn

*Rc , continuous rice; ReS, rice-soybean; R� PS , rice-short-term pasture; R� PL , rice-long-Trifolium alejandrinum. Short term pasture, mix of: Trifolium pratense and Lolium multiflcorniculatus.

2.4. Estimate forage for grazing

Perennial R� PS and R� PLpastures were rotational grazed withsheep 7e8 times a year for each plot, although meat productionwas not quantified because the 7-10-day grazing periods were tooshort. Therefore, beef cattle productionwas estimated as a functionof measured forage production, assuming variable use (40e70%)depending on age and growing season in R� PL, and 60% utilizationin R� PS(Rovira et al., 2009). In this study was assumed that in R�PS, 14 kg dry weight (DW) pasture was needed to produce 1 kg ofmeat and that for R� PL, 12e14 kg DW of grass was needed toproduce 1 kg of meat (NRC, 2000). In addition, it was assumed astocking rate of two calves per ha, with a live weight (LW) of 160 kgeach, in R� PS. After 1.5 years, these animals each weighed 290 kgLW. In R� PL, each animal-starting at 160 kg LW-achieved 537 kgafter 2.5 years, after which two more animals entered and after 1year weighed 293 kg each.

2.5. Energy performance

The energy performance scope assessed herein followed a lifecycle assessment approach: the life cycle inventory (LCI) includedinputs and outputs from the cradle to the farm gate (Table 4). Theenergies involved in the transport of inputs, products, manual la-bor, solar radiation and the biological fixation of nitrogen were notconsidered. The functional unit of this study was the EP from the EIof each crop rotation by year and per ha. For this reason, each itemlisted by the LCI was converted into equivalent energy unitsexpressed in MJ. As EI and EP were expressed in MJ ðha yrÞ�1, theEROI - the ratio between EP and EI - was expressed in MJ MJ�1

(Table 4). The LCI considered all inputs and outputs of sowing,postplanting (fertilization, application of pesticides), and harvestoperations. Input data were from the following sources: INIA LTE-

-winter).

term pasture. Rice, Oryza sativa L. Soybean, Glycinemax L. Lm, Loliummultiflorum. Ta,orum. Long-term pasture, mix of: Festuca arundinacea, Trifolium repens, and Lotus

Table 3Fertilizers and pesticide (kg ha�1) use (and frequency) applied, based on crop and pasture cycle in each rotation system.

Rc ReS R� PS ReP

rice rice soybean Rice pasture 1st rice 2nd rice pasture pasture pasture

yr 1 yr 2 yr 3

Fertilizers Triple super phosphate) 100(1) 165(1)

(kg ha�1) (0e46/46-0)N-ðPsol =PtotalÞ-K Diammonium phosphate 130(1)

(18e46/46-0)Potassium chloride 60(1) 80(1) 50(1) 95(1) 27(1) 48(1)

(0-0/0-60)Chemical synthesized 60(1) 60(1) 60(1) 60(1) 60(1) 60(1)

(9e25/25-25)Physical synthesized 150(1) 150(1)

(15e30/30-15)Phosphate rock 200(1) 200(1)

(0e12/23-0)Urea 150(2) 130 þ 30(1) 86 þ 117(1) 90 þ 70(1) 100 þ 70(1)

(46e0/0-0)*Herbicides Glyphosate 1.68(2) 1.68(2) 1.68(3) 1.68(2) 1.92(1) 1.68(2) 1.68(2) 1.92(1)

(kg of a.i.ha�1) 2,4-D 0.72(1) 0.72(1) 0.72(1) 0.72(1)

Imazapyr þ Imazapic 0.098(2)

Fluroxypyr 0.1(1) 0.1(1) 0.1(1) 0.1(1)

Clomazone 0.4(1) 0.4(1) 0.4(1) 0.4(1)

Quinclorac 0.38(1) 0.38(1) 0.38(1) 0.38(1)

Pyrazosulfuron 0.04(1) 0.04(1) 0.04(1) 0.04(1)

Penoxsulam 0.64(1) 0.64(1) 0.64(1) 0.64(1)

Propaquizafop 0.15(1)

Metolachlor 0.96(1)

Flumetsulam 0.06(1) 0.06(1)

Clethodim 0.168(1)

Fungicides Azoxystrobin þ 0.2(1) 0.2(1) 0.2(1) 0.2(1) 0.2(1) 0.2(1)

(kg of a.i. ha�1) kresoxim-methyl þciproconazoleAzoxystrobin þ 0.175(1)

ciproconazoleInsecticides Trichlorfon 0.8(1)

(kg of a.i.ha�1) Imidacloprid þ 0.115(1)

beta-cyfluthrin

* Urea split in all rice crops: at tillering and panicle initiation stage; a:i:: active ingredient. Rice and soybean seeds were treated with difenoconazole (0.00006 kg kg�1 of seed)and thiamethoxam (0.00035 kg kg�1 of seed).** NePeK expressed as percentage.

Table 4Energy conversion factors used for different inputs used in the studied rice-cropping systems.

LCI inputs Energy conversion factor References

Seeds (MJ kg�1)Rice 17.6 Determined by the studySoybean 23.50 Determined by the studyForage legume 17.20 Muhammad et al. (2014)Forage gramineae 36.10 Fuksa et al. (2013)

Fertilizer (MJ kg�1)Nitrogen 51.47 Hill et al. (2006)Phosphorus 9.17 Hill et al. (2006)Potassium 5.96 Hill et al. (2006)

Fuels (MJ L�1) FuelsDiesel 43.99 Nagy (1999)Aviation gasoline 34.78 Petroleum Products Division (2016)

Pesticides (MJ kg�1)Glyphosate 454.00 Audsley et al. (2014)2-4 D amine 87.00 Audsley et al. (2014)Fluroxypyr-meptyl 518 Audsley et al. (2014)Herbicides 303.80 Pimentel and Pimentel (2007)Insecticides 418.40 Pimentel and Pimentel (2007)Fungicides 115.00 Pishgar-Komleh et al. (2011)

I. Macedo et al. / Journal of Cleaner Production 278 (2021) 1237714

RC records; the Uruguayan Association of Agricultural Services(CUSA, 2017); and personal interviews with companies that con-ducted specific operations (e.g. aerial fertilizations). The LCI of rice-cropping systems was organized into seven categories: irrigation;

fuel; machinery; phosphorus and potassium fertilization; nitrogenfertilization; pesticides (herbicides, insecticides, and fungicides);and others (seed, cattle). The energy produced, expressed inMJ ha�1, considered both the grain yield of rice (paddy) and

Fig. 2. Box plot of rice and soybean grain yields and forage production of pasture crops,and potential beef production in 2015e2017 in different rotation systems. Yieldsconsidered a dry weight with 13% humidity. Grain yield and forage are scaled to the leftaxis, and beef production is scaled to the right axis. Grain yield is expressed as kg ha-1and forage and meat as kg ðha yrÞ�1.

I. Macedo et al. / Journal of Cleaner Production 278 (2021) 123771 5

soybean crops and the potential meat production of pasture pro-duction in those rotations. For all products were estimated theirenergy equivalent using the corresponding calorific power: 19.6MJ kg�1 for forage (Portugal-Pereira et al., 2015) and 9.3 MJ kg�1

for meat (Restle et al., 2001). The calorific power value of the paddyrice and the soybean grain was determined according to ASTM D4442-07 standard, “Standard Test Methods for Direct MoistureContent Measurement of Wood and Wood-Base Materials, MethodA” (Table 4).

2.6. Statistical analysis

Productivity of each crop was compared, using single-sample ttests, to the official national mean annual yield record for: rice,8390 kg ha�1; soybean, 2400 kg ha�1; pastures, 5500 kg ha�1; andbovine cattle beef 148 kg ha�1 (DIEA, 2018; Simeone et al., 2008).Mean EP and EROI comparisons between crop rotation, all phases(rice, soybean, and pasture/cattle), and rice-only crops in differentrotations systems were made using Fisher’s ANOVA tests whenparametric assumptions (normality and homoscedasticity) weresatisfied, with Tukey’s honest significant difference tests for posthoc analyses. Otherwise, comparisons were made using post hocKruskal-Wallis and Dunn’s tests. All statistical tests were run withthe statistical package R version 2.12.0 (R Development Core Team,2018) for the platform i686-pc-Linux-gnu (64-bit) with R Com-mander 1.5e4 (Fox, 2011; R Core Team, 2013) on a GNU/Linuxoperating system.

3. Results

3.1. Biomass production

Rice experimental yield (10,147 kg ha�1) was 21% higher thanthe commercial mean yield of the corresponding years. The lowestrice productivity was found in RC (9741 kg ha�1) and the highestwas the first rice in R� PL (11,043 kg ha�1).

Mean soy yield at the site was 2868 kg ha�1, which is 20%greater than the national average of 2400 kg ha�1 reported by DIEA(2018). Annual pasture biomass production ranged between 4240and 7337 DM ha�1, with a mean value of 5897 kg ha�1. Pastureproductivity was higher in R� PL (p-value ¼ 0.0099) comparedwith R� PS, which is comparable with commercial farm valuesdescribed by Simeone et al. (2008). Finally, annual estimated beefproduction was 257 kg ha�1, which is higher (p-value<0.0001)than commercial productivity reported by Simeone et al. (2008)(Fig. 2).

3.2. Crop rotation energy efficiencies

The ReS rotation’s EROI was 11% higher than that of R� PL, butno improvements were found in other rotations. However, whenanimal production was excluded, the EROI of R� PLand R� PS was74% and 36% greater, respectively (Table 5).

Rice rotations with perennial pastures (R� PL and R� PS) had44% lower EP compared with continuous cropping systems (ReSand RC). The RC EP was 231% higher than that of R� PL. However,when animal production was excluded, the EP of these systemsmatched those of the ReS rotation. Conversely, RC had a 246%higher EI than R� PL (Table 5).

Three components explained approximately 60% of the EI in allsystems: irrigation, fuel use and nitrogen fertilization (Fig. 3). In theReS system, the EI in pesticides (17%) and phosphate and potas-sium fertilization (7%) had higher relative weights compared withother systems in which these components were 6e10% (pesticides)and about 4% (P and K fertilization). Nitrogen fertilization explained

33% of the EI in RC , while in other systems its relative weight was14e22% of EI.

3.3. Energy balance in the rotation phases

Examining the rotation phases, soybean and rice crops recordedsimilar EROI values (~7 MJ MJ�1), except in RCwhich had 5.7MJ MJ�1 and differed only from the first rice in R� PL. The EROIvalues of pasture phases when animal production was includedwere the lowest among all phases. However, when animal pro-duction was excluded, and only forage production was accountedfor, the values obtained by pastures were the highest, particularly inR� PL (Table 6).

The EP for soybean crops was 6% lower than the average for allrice phases (155,374 MJ ðha yrÞ�1). When animal production wasexcluded from the pasture phases, EP increased 15 and 17 times forR� PS and R� PL, respectively, with a similar EP for soybeans.Soybean and pasture phases had 64% and 77% lower EI, respec-tively, than corresponding rice phases (Table 6).

Irrigation, fuel, and nitrogen fertilization accounted forapproximately 70% of rice phase EI (Fig. 4). Both pastures andsoybean phases had low EI demands, related to nitrogen fertiliza-tion. However, the EI for phosphorus and potassium fertilizationwas 8e14%, which is significantly higher than the EI in rice phases,at 2e4%. Pesticides represented 38% of the EI in the soybean phasein ReS, while it varied for other phases from 7% to 14%. In thepasture phases, other components (e.g. animals, seeds) becamemore relevant and represented 63% and 55% in R� PL and R� PS,respectively.

EI related to the depreciation of machinery was irrelevant anddid not exceed 1% in any of the phases evaluated.

3.4. Rice crop energy balance in different rotation systems

The highest EROI was found in the first rice following long-termpastures (R� PL) and in rice in rotationwith soybeans (ReS). Rice inrotationwith short-term pastures did not differ comparedwith ricerotated with soybean or with the second rice in R� PL. Finally, EROI

Table 5Energy information expressed in MJ ðha yrÞ�1 for mean value and standard deviation of invested energy; produced energy, with (þA) and without (-A) beef production of:continuous rice with pasture during the winter (RC ) (n ¼ 6); rice-soybean (R� S, 2 years) (n ¼ 6); rice, followed by short pastures (R� PS , 2 years) (n ¼ 6) and 1st rice�2ndricefollowed by a three and a half year pasture (R� PL , 5 years) (n ¼ 6). The EROI is expressed in MJ MJ�1.

Crop rotation EP EI EROIþ A -A þ A -A

R� PL 64,540 ± 2309 D 106,361 ± 2901 b 10,607 6.1 B 10.0 a

R� PS 80,697 ± 6117 C 109,010 ± 10,135 b 14,500 5.6 C 7.5 b

ReS 109,803 ± 11,279 B 109,803 ± 11,279 b 15,153 7.2A 7.2 b

RC 149,158 ± 8765 A 149,158 ± 8765 a 26,117 5.7 BC 5.7 c

Values followed by the different letter are significant different for a P<0.05.

Fig. 3. Distribution of energy invested in four rice production systems, expressed inpercentage units. Continuous rice (RC , black dots); rice-soybean (ReS, black solid line);rice with short-term pastures (R� PS , black dashed line), and first to second rice fol-lowed by long-term pasture (R� PL , grey solid line).

I. Macedo et al. / Journal of Cleaner Production 278 (2021) 1237716

estimated in rice monoculture was 21% lower than the average ofthe rice crops in rotation and showed a pattern similar to the sec-ond rice in R� PL (Fig. 5). There were no large variations in rice EPbetween rotations. On average, EP was 155,374 MJ ha�1; themaximum value was found for the first rice in R� PL, which was9.6% higher than the second rice in the same rotation (Fig. 5). Thetwo rice crops seeded on rice residues (second rice in R� PL and RC)had the lowest EP, different from the bedside rice in R� PL and therice in ReS (Fig. 5). The first rice in R� PL had the lowest EI. Thesecond rice in R� PL, the rice in rotation with short-term pastures,and with soybeans had EI values increased by 1%, 6%, and 3%,respectively, compared with the first rice in R� PL. The rice in RChad a 23% higher EI than the first rice in R� PL. Energy consumptionby irrigation, fuel, and nitrogen fertilization accounted forapproximately 70% of the EI in rice cropping in all rotations. The EIin nitrogen fertilization was higher in rice monoculture comparedwith other rice crops in rotation, representing 33% of the EI.

Table 6Energy information expressed inMJ ðha yrÞ�1 for mean value ± standard deviation (numbbeef production of crops used in crop rotations: continuous rice with pasture during theyears) and 1st rice� 2ndrice followed by a three and a half year pasture (R� PL , 5 years).

Crop rotation Crop EP

þA

R� PL 1stRice 161,809 ± 6986 (6) A

2ndrice 147,583 ± 16,981 (6) AB

Pasture 4435 ± 279 (18) CD

R� PS Rice 157,353 ± 12,120 (6) BC

Pasture 4040 ± 466 (6) D

ReS Rice 160,967 ± 8179 (6) A

Soybean 58,640 ± 16,295 (6) BC

RC Rice 149,158 ± 8765 (6) AB

Values followed by the different letter are significant different for a P<0.05.

Meanwhile, phosphorus and potassium fertilization EI in rice inReS and RC was 4%, while in rice in rotation with pastures (R� PLand R� PS) was 2% (Fig. 6).

4. Discussion

Conservation management practices including diversifiedfarming systems are critical to reach an ecological intensification offood production in agroecosystems (Cassman, 1999). However,intensification pathways do not always fit with ecological intensi-fication principles (Zhao et al., 2009), especially when the agro-ecosystem has a high dependence of inputs such as fertilizers,pesticides, and fossil energy (Godfray et al., 2010; Pimentel, 2009;Tilman et al., 2011). The challenge of intensified agriculture implyan efficient, smart and strategic use of inputs (e.g. pesticides, fer-tilizers, etc.), minimizing emissions (e.g. green-house gases, waterpollution, etc.) increasing natural capital (e.g. soil organic carbon,soil structure), diversifying production and promoting environ-mental services that strengthening resilience as purpose byConway (2012); FAO (2009); Soemarwoto and Conway (1992). Thisgoal to request go beyond the Green Revolution principles, becausein addition to increase productivity, in a lower yield gap situation,needs a holistic and broad assesment. Then, it is necessary to selectspecific management practices that may allow our agroecosystemsto be adjusted, in the best possible way, with their intrinsic char-acteristics (for example, incidence of weeds, pests and diseases, soilfertility and climate change). Examples of this are the JavaneseHome Garden, in several tropic regions for rice production(Soemarwoto and Conway, 1992); the rice-fish co-culture systems(Wan et al., 2018, 2019b; a, 2020) and the integrated rice-beefsystems observed at Temperate Grassland terrestrial ecoregion(South America) (Denardin et al., 2019; Olson et al., 2001). How-ever, in the rice-livestock systems, there are some questions stillremaining in the designing and fine tuning of the prevalent rota-tions and their sustainability. Rice rotation systems analyzed in thispaper, with different proportion and length of pastures and other

er of replicates) of invested energy and produced energy, with (þA) and without (-A)winter (RC ); rice-soybean (ReS, 2 years); rice, followed by short pastures (R� PS , 2The EROI is expresed in MJ MJ�1.

EI EROI

-A þA -A

161,809 ± 6986 (6) a 21,183 7.6 A 7.6 b

147,583 ± 16,981 (6) a 21,134 7.0 AB 7.0 bc

74,137 ± 7577 (18) b 3573 1.2 CD 20.7 a

157,353 ± 12,120 (6) a 22,23 7.1 AB 7.1 bc

60,666 ± 13,752 (6) b 6771 0.6 D 9.0 b

160,967 ± 8179 (6) a 22,276 7.2 AB 7.2 bc

58,640 ± 16,295 (6) b 8031 7.3 AB 7.3 bc

149,158 ± 8765 (6) a 26,117 5.7 BC 5.7 c

Fig. 4. Percentage distribution of the invested energy according to crop rotation phase: two rice crops followed by long-term pasture (A), rice with short-term pasture (B), rice-soybean (C), and continuous rice (D).

I. Macedo et al. / Journal of Cleaner Production 278 (2021) 123771 7

crops, are representative of the diversity and natural conditions ofecosystems from Temperate Grassland terrestrial ecoregion (Olsonet al., 2001).

In the current study, intensification of a rice-pasture system,including a shorter-term pasture or replacing it with other crops,decreased the system’s energy efficiency if animal production wasnot considered. These results are consistent with studies showingthat less intense systems are more energy-efficient (Hülsbergenet al., 2001; Rathke et al., 2007). Similar to reports by Borin et al.(1997) and Kuesters and Lammel (1999), the results showed thatincrease in EI increased EP. However, this was nonlinear: energyefficiencywas reduced by intensifying the rotation. Although the RCrotation had the highest EP, its efficiency was lower than other ricesystems in rotation with pastures or soybeans. Similar results werefound by Franzluebbers and Francis (1995) with sorghum and cornand by Rathke et al. (2007) with soybeans and corn in monocultureor in rotation. Pimentel (2019) found that nitrogen fertilizationrepresented 16e35% of the total EI in corn production. The nitrogenfertilizationwas one of the most important components explainingEI in this study, ranging from14% to 33% in ReS and RC , respectively.In the most intensive systems, the increase in EI associated withnitrogen fertilizer was higher than the increase in EP. Franzluebbersand Francis (1995) also observed a reduction in energy efficiencywith the addition of nitrogen. Conversely, Theisen et al. (2017)

reported improvements in energy efficiency by intensifying arotation of three years of rice and three fallow years (with grazing),with the inclusion of soybeans or pastures. The results of this workshowed similar results when animal productionwas considered; inthat case, only ReS system improved energy efficiency. The EI re-ported by Theisen et al. (2017) was higher than shown in therotation systems evaluated in this study, which is likely due todifferences in nitrogen fertilization (71 vs. 53 kg ðha yrÞ�1). How-ever, the EP reported by Theisen et al. (2017) was 27% lower thanthe EP observed herein; this is likely explained by their 26% lowerrice productivity comparedwith the current study (10,147 kg ha�1).Another important component explaining EI is the energy used topump water for rice flooding. Rotation phases with rice had 264%more EI than rotation phases with pastures or soybean, which werenot irrigated.Franzluebbers and Francis (1995), reported that irri-gated maize and sorghum systems have energy costs that are200e300% higher than the same systems under rain-fed conditionsand, thus, lower energy efficiency. Although the current studycompared different crops/phases, rice crops had similar energyefficiency to dry land soybean cropping (7.3 MJ MJ�1). The EI inpesticides in the soybean phase was 38%. Ranges of 17e20% and7e31% have been cited for pesticide EI in crops by Swanton et al.(1996) and Rathke et al. (2007), respectively. For the presentstudy, the EI values for meat production during the pasture phases

Fig. 5. Energy information for rice cultivation in different rice cropping systems: continuous rice (RC ); rice-soybean (ReS); rice and short-term pastures (R� PS), and first rice cropfollowed by second rice crop and long-term pasture (R� PL). EI: energy input; EP: produced energy; EROI: energy return on investment.

Fig. 6. Percentage of invested energy in each rice crop in different crop rotationsystems.

I. Macedo et al. / Journal of Cleaner Production 278 (2021) 1237718

were 7.5 and 12.5 MJ kg LW�1 in R� PLand R� PS, respectively.Similar results have been reported by Modernel et al. (2013) insown pastures.

Pasture longevity in R� PL, allowed for some components of theEI seeding costs (fuel, fertilization, and pesticides) to be diluted bygreater time compared with the pasture of R� PS. The mean energyefficiency in the rice cropping observed in this work was approxi-mately 7 MJ MJ�1. Another rice study using 20 years of data fromUruguayan farmers showed similar results (6.88 MJ MJ�1, EI of17,000 MJ ha�1, and net energy of 100,000 MJ ha�1) for the end ofthe same period (2009e2013) (Pittelkow et al., 2016). Anotherstudy in Vietnam obtained similar efficiencies but calculated EP asthe energy contained in both the stubble and grain (Truong et al.,2017). Truong et al. (2017) reported rice grain EP of74,526e82,964 MJ ha�1, values that are lower than the mean EP of155,368 MJ ha�1 observed herein. However, rice productivity in

Truong et al. (2017) was 5e5.6Mg ha�1 compared with 10Mg ha�1

obtained in the current experiment.The EI values for rice in rotation with pastures or soybean (ReS,

R� PS, and R� PL) were lower than rice in monoculture and similarto the values of 17,000e20,000 MJ ha�1 reported by (Pittelkowet al., 2016). In contrast, EI of 25,000e34,000 MJ ha�1 was re-ported in high intensity rice systems of Central China (Yuan andPeng, 2017) and Vietnam (Truong et al., 2017), which are similarto those obtained in the Rc system (25,700MJ ha�1). This indicatesa greater input dependence in high-intensity systems, which ismainly explained by the higher nitrogen fertilization of 1900e5900MJ ha�1 Yuan and Peng (2017) and 1900e8700 MJ ha�1 (Truonget al., 2017), consistent with the values reported in this work. Inthe present study, the nitrogen fertilization EI in monoculture rice(8600 MJ ha�1) was 95% higher than the EI in rice rotating withpastures or soybeans. Similarly, irrigation represented a higherproportion (40%) of the total EI in rice cropping compared with15e20% of the total EI reported by Pittelkow et al. (2016). Thesedifferences can be explained by their predominant use of electricrather than diesel pumping.

5. Conclusions

This study showed that the highest energy production occurs inmore intense rice systems, which use inputs that required higherenergy levels. This must be balanced in the long term, whennonrenewable energy resources are finite. This study showed that,without considering animal production, the systems with thehighest energy efficiency was rice in rotation with long-termperennial pastures.

Intensification alternatives exist for rice cropping systems,which improve energy efficiency compared with the rice-long-term pasture rotations. The rotation of rice with soybeansimproved the energy return on investment compared with rotationof rice with long-term pastures, when animal production wasaccounted for. In any case, rice-pasture rotation consumed lessenergy, which makes it more sustainable. In addition, rice cropsthat rotated with either soybeans or pasture required less energy

I. Macedo et al. / Journal of Cleaner Production 278 (2021) 123771 9

investment and achieved better energy use efficiency than ricemonoculture.

From this perspective, the results of this study elucidate thatincreasing pasture participation in cropping systems play a signif-icant role in achieving sustainable cropping rotation systems. Forthe Temperate Grassland Terrestrial Ecoregion of South America,EROI assessments of four business as usual rice production systemsallowed to discriminate and hierarchize their sustainability anddiversity.

Future studies should include other indicators to evaluateecological intensification. They should also identify potential trade-offs between sustainability dimensions. This work focused on oneaspect of environmental sustainability but did not assess othersustainability dimensions. Further works should use a holistic, in-tegrated approach and potentially include carbon, water, and nu-trients footprints.

CRediT authorship contribution statement

Ignacio Macedo: Conceptualization, Methodology, Validation,Formal analysis, Investigation. Jos�e A. Terra: Conceptualization,Methodology, Writing - review & editing, Project administration,Supervision. Guillermo Siri-Prieto: Conceptualization, Writing -review & editing, Supervision. Jos�e Ignacio Velazco: Methodology,Resources. Leonidas Carrasco-Letelier: Conceptualization, Meth-odology, Writing - review & editing, Visualization.

Declaration of competing interest

The authors declare that they have no known competingfinancial interests or personal relationships that could haveappeared to influence the work reported in this paper.

Acknowledgments

We thank the National Institute of Agricultural Research (INIA)and the National Research and Innovation Agency (ANII) for fund-ing this work through a national postgraduate scholarship (POSNAC 2015 1 109776).

Appendix A. Supplementary data

Supplementary data related to this article can be found athttps://doi.org/10.1016/j.jclepro.2020.123771.

References

Audsley, E., Stacey, K., Parsons, D.J., Williams, A.G., 2014. Estimation of the Green-house Gas Emissions from Agricultural Pesticide Manufacture and Use. Cran-field University, pp. 1e20.

Bao, L., Ginella, J., Cadenazzi, M., Castiglioni, E.A., Martínez, S., Casales, L.,Caraballo, M.P., Laborda, �A., Simo, M., 2018. Spider assemblages associated withdifferent crop stages of irrigated rice agroecosystems from eastern Uruguay.Biodivers. Data J. 6, e24974.

Bird, G.W., Edens, T., Drummond, F., Groden, E., 1990. Design of pest managementsystems for sustainable agriculture. In: Sustainable Agriculture in TemperateZones, pp. 56e57.

Borin, M., Menini, C., Sartori, L., 1997. Effects of tillage systems on energy and carbonbalance in north-eastern Italy. Soil Tillage Res. 40 (3e4), 209e226.

Bryant, R.J., Anders, M., McClung, A., 2012. Impact of production practices onphysicochemical properties of rice grain quality. J. Sci. Food Agric. 92 (3),564e569.

Carracelas, G., Hornbuckle, J., Rosas, J., Roel, A., 2019. Irrigation management stra-tegies to increase water productivity in Oryza sativa (rice) in Uruguay. Agric.Water Manag. 222, 161e172.

Carvalho, P.C.d.F., Barro, R.S., Neto, A.B., Nunes, P.A.d.A., de Moraes, A., Anghinoni, I.,Bredemeier, C., Bayer, C., Martins, A.P., Kunrath, T.R., dos Santos, D.T.,Carmona, F.d.C., Barros, T., Filho, W.d.S., de Almeida, G.M., Caetano, L.A.M.,Cecagno, D., Arnuti, F., Denardin, L.G.d.O., Bonetti, J.d.A., de Toni, C.A.G.,Borin, J.B.M., 2018. Integrating the pastoral component in agricultural systems.

Rev. Bras. Zootec. 47.Cassman, K.G., 1999. Ecological intensification of cereal production systems: yield

potential, soil quality, and precision agriculture. Proc. Natl. Acad. Sci. Unit. StatesAm. 96 (11), 5952e5959.

Castillo, J., Terra, J., M�endez, R., 2015. Nitrogen fertilization recommendation on ricebased on indicators (Indicadores objetivos para la recomendaci�on de fertil-izaci�on N en arroz). In: XII Conferencia Internacional de Arroz para Am�ericaLatina e Caribe. Porto Allegre, p. 125 (In Spanish). URL. http://www.ainfo.inia.uy/digital/bitstream/item/5674/1/Congreso-2015-Castillo-p.125.pdfcastilloetal2014.

Conforti, P., Giampietro, M., 1997. Fossil energy use in agriculture: an internationalcomparison. Agric. Ecosyst. Environ. 65 (3), 231e243.

Conway, G., 2012. One Billion Hungry: Can We Feed the World. Cornell UniversityPress.

Cusa, 2017. Uruguayan Chamber of Agricultural Services (C�amara Uruguaya deServicios Agropecuarios) (In Spanish). URL. http://www.cusa.org.uy/cusa/precios{_}servicios{_}agricolas.

de Moraes, A., de Faccio Carvalho, P.C., Anghinoni, I., Lustosa, S.B.C., deAndrade, S.E.V.G., Kunrath, T.R., et al., 2014. Integrated cropelivestock systemsin the Brazilian subtropics. Eur. J. Agron. 57, 4e9.

Deambrosi, E., 2003. Rice production system in Uruguay and its sustainability. URL.http://www.ainfo.inia.uy/digital/bitstream/item/13278/1/conf-clima-temp-3p.19.pdf.

Deambrosi, E., M�endez, R., Castillo, J., 2015. Elements to consider in the inclusion ofpotassium in rice fertilization (Elementos a considerar en la inclusi�on depotasio en la fertilizaci�on del arroz) (In Spanish).

Denardin, L.G.d., Carmona, F.d., Veloso, M.G., Martins, A.P., Freitas, T.F. S.d.,Carlos, F.S., Marcolin, E., Camargo, F.A.d.O., Anghinoni, I., Mar. 2019. No-tillageincreases irrigated rice yield through soil quality improvement along time. SoilTillage Res. 186, 64e69.

Díaz-Zorita, M., Duarte, G.A., Grove, J.H., 2002. A review of no-till systems and soilmanagement for sustainable crop production in the subhumid and semiaridPampas of Argentina. Soil Tillage Res. 65 (1), 1e18.

DIEA, 2018. Agricultural Statistic Yearbook 2018 (Anuario Estadístico Agropecuario2018). Ministerio de Agricultura, Ganadería y Pesca, Editorial Hemisferio Sur (InSpanish). URL. https://descargas.mgap.gub.uy/DIEA/Anuarios/Anuario2018/Anuario_2018.pdf.

Dur�an, A., Califra, A., Molfino, J.H., Lynn, W., 2006. Keys to Soil Taxonomy forUruguay. US Department of Agriculture, Natural Resources Conservation Ser-vice, Washington, DC.

FAO, 2009. World agriculture in the perspective of the year 2050 (La agriculturamundial en la perspectiva del a~no 2050) (In Spanish). URL. http://www.fao.org/fileadmin/templates/wsfs/docs/I.

Fox, J., 2011. R commander. URL. http://socserv.mcmaster.ca/jfox/Misc/Rcmdr/.Franzluebbers, A.J., Francis, C.A., 1995. Energy output: input ratio of maize and

sorghum management systems in eastern Nebraska. Agric. Ecosyst. Environ. 53(3), 271e278.

Fuksa, P., Hakl, J., Brant, V., 2013. Energy balance of catch crops production. Zem-dirbyste-Agriculture 100 (4), 355e362. Others.

García-Pr�echac, F., Ernst, O., Siri-Prieto, G., Terra, J.A., 2004. Integrating no-till intocropepasture rotations in Uruguay. Soil Tillage Res. 77 (1), 1e13.

Godfray, H.C.J., Beddington, J.R., Crute, I.R., Haddad, L., Lawrence, D., Muir, J.F.,Pretty, J., Robinson, S., Thomas, S.M., Toulmin, C., 2010. Food security: thechallenge of feeding 9 billion people. Science 327 (5967), 812e818.

Heichel, G., Barnes, D., 1984. Chapter 4: opprotunities for meeting crop nitrogenneeds from symbiotic nitrogen fixation. In: Organic Farming: Current Tech-nology and its Role in a Sustainable Agriculture. American Society of Agronomy,Crop Science Society of America, Soil Science Society of America, pp. 49e59.

Hern�andez, J., Berger, A., Deambrosi, E., Lavecchia, A., 2013. Soil phosphorus tests forflooded rice grown in contrasting soils and cropping history. Commun. Soil Sci.Plant Anal. 44 (7), 1193e1210.

Hill, J., Nelson, E., Tilman, D., Polasky, S., Tiffany, D., 2006. Environmental, economic,and energetic costs and benefits of biodiesel and ethanol biofuels. Proc. Natl.Acad. Sci. Unit. States Am. 103 (30), 11206e11210.

Hülsbergen, K.-J., Feil, B., Biermann, S., Rathke, G.-W., Kalk, W.-D., Diepenbrock, W.,2001. A method of energy balancing in crop production and its application in along-term fertilizer trial. Agric. Ecosyst. Environ. 86 (3), 303e321.

Kuesters, J., Lammel, J., 1999. Investigations of the energy efficiency of the pro-duction of winter wheat and sugar beet in Europe. Eur. J. Agron. 11 (1), 35e43.

Liebman, M., Janke, R., 1990. Sustainable weed management practices. SustainableAgriculture in Temperate Zones. John Wiley and Sons, pp. 111e143.

Maraseni, T., Chen, G., Banhazi, T., Bundschuh, J., Yusaf, T., 2015. An assessment ofdirect on-farm energy use for high value grain crops grown under differentfarming practices in Australia. Energies 8 (11), 13033e13046.

Martins, A.P., Denardin, L.G.d.O., Borin, J.B.M., Carlos, F.S., Barros, T., Oz�orio, D.V.B.,Carmona, F.d.C., Anghinoni, I., Camargo, F.A.d.O., Carvalho, P.C.d.F., 2016. Short-term impacts on soil-quality assessment in alternative land uses of traditionalpaddy fields in southern Brazil. Land Degrad. Dev. 28 (2), 534e542.

Modernel, P., Astigarraga, L., Picasso, V., 2013. Global versus local environmentalimpacts of grazing and confined beef production systems. Environ. Res. Lett. 8(3), 035052.

Muhammad, D., Misri, B., El-Nahrawy, M., Khan, S., Serkan, A., 2014. Egyptian Clover(Trifolium Alexandrinum); King of Forage Crops. Regional 0ffice for the Near Eastand North Africa. Food and Agriculture Organization of the United Nations,Cairo. URL. http://www.fao.org/3/a-i3500e.pdf.

I. Macedo et al. / Journal of Cleaner Production 278 (2021) 12377110

Nagy, C.N., 1999. Energy Coefficients for Agriculture Inputs in Western Canada.Centre for Studies in Agriculture, Law and the Environment. University ofSaskatchewan.

NRC, 2000. Nutrient Requirements of Beef Cattle: 7th Rev. Update 2000.Olson, D.M., Dinerstein, E., Wikramanayake, E.D., Burgess, N.D., Powell, G.V.N.,

Underwood, E.C., D’amico, J.A., Itoua, I., Strand, H.E., Morrison, J.C., Loucks, C.J.,Allnutt, T.F., Ricketts, T.H., Kura, Y., Lamoreux, J.F., Wettengel, W.W., Hedao, P.,Kassem, K.R., 2001. Terrestrial Ecoregions of the World: a New Map of Life onEarthA new global map of terrestrial ecoregions provides an innovative tool forconserving biodiversity. Bioscience 51 (11), 933e938.

Patterson, H.D., 1964. Theory of cyclic rotation experiments. J. Roy. Stat. Soc. B 1e45.Pellegrini, P., Fern�andez, R.J., 2018. Crop intensification, land use, and on-farm en-

ergy-use efficiency during the worldwide spread of the green revolution. Proc.Natl. Acad. Sci. Unit. States Am. 115 (10), 2335e2340.

Petroleum Products Division, 2016. Chapter: 13 Characteristics of Petroleum Prod-ucts Stored and Dispensed.

Pimentel, D., 2009. Energy inputs in food crop production in developing anddeveloped nations. Energies 2 (1), 1e24.

Pimentel, D., 2019. Handbook of Energy Utilization in Agriculture. CRC press.Pimentel, D., Pimentel, M.H., 2007. Food, Energy, and Society. CRC press.Pishgar-Komleh, S.H., Sefeedpari, P., Rafiee, S., 2011. Energy and economic analysis

of rice production under different farm levels in Guilan province of Iran. Energy36 (10), 5824e5831.

Pittelkow, C.M., Zorrilla, G., Terra, J., Riccetto, S., Macedo, I., Bonilla, C., Roel, A., 2016.Sustainability of rice intensification in Uruguay from 1993 to 2013. Global FoodSecurity 9, 10e18.

Portugal-Pereira, J., Soria, R., Rathmann, R., Schaeffer, R., Szklo, A., 2015. Agriculturaland agro-industrial residues-to-energy: techno-economic and environmentalassessment in Brazil. Biomass Bioenergy 81, 521e533.

R Core Team, 2013. R: A Language and Environment for Statistical Computing. RFoundation for Statistical Computing. URL. http://www.R-project.org/.

Rathke, G.-W., Wienhold, B.J., Wilhelm, W.W., Diepenbrock, W., 2007. Tillage androtation effect on cornesoybean energy balances in eastern Nebraska. SoilTillage Res. 97 (1), 60e70.

Restle, J., Cerd�otes, L., Vaz, F.N., Brondani, I.L., 2001. Carcass and meat characteristicsof charolais and 3/4 charoles 1/4 nellore heifers, finished in feedlot (Caracter-ísticas de carcaça e da carne de novilhas Charoles e 3/4 Charoles 1/4 Nelore,terminadas em confinamento). Rev. Bras. Zootec. 30 (3), 1065e1075 (InPortuguese).

Rovira, P.J., Velazco, J.I., Bonilla, O., 2009. Effect of the supplementation strategy(Self-consumption vs. infrequent supplementation) and the level of salt in thesupplement on the productive performance of steers ending during the sum-mer (Efecto de la estrategia de suplementaci�on (Autoconsumo vs.Suplementaci�on infrecuente) y el nivel de sal en el suplemento sobre eldesempe~no productivo de novillos en terminaci�on durante el verano). In:Deambrosi, E., Montossi, F., Saravia, H., Blanco, P.H., Ayala, W. (Eds.), 10 a~nos dela Unidad de Producci�on Arroz-Ganadería, first ed. INIA, Montevideo,

pp. 99e104 (In Spanish). URL. http://www.ainfo.inia.uy/digital/bitstream/item/7607/1/st-180-p.99-104.pdf.

Seck, P.A., Diagne, A., Mohanty, S., Wopereis, M.C., 2012. Crops that feed the world 7:Rice. Food security 4 (1), 7e24.

Simeone, A., Andregnette, B., Buffa, J.I., 2008. Efficient meat production in rice-pasture systems (Producci�on de carne eficiente en sistemas arroz-pasturas).INIA. Serie FPTA (In Spanish) (022). URL. http://www.ainfo.inia.uy/digital/bitstream/item/12654/1/18429130709152516.pdf.

Soemarwoto, O., Conway, G.R., 1992. The Javanese homegarden. J. farming Syst.research-extension 2 (3), 95e118.

Swanton, C.J., Murphy, S.D., Hume, D.J., Clements, D.R., 1996. Recent improvementsin the energy efficiency of agriculture: case studies from Ontario, Canada. Agric.Syst. 52 (4), 399e418.

Theisen, G., Silva, J.J.C., Silva, J.S., Andres, A., Anten, N.P.R., Bastiaans, L., 2017. Thebirth of a new cropping system: towards sustainability in the sub-tropicallowland agriculture. Field Crop. Res. 212, 82e94.

Tilman, D., Balzer, C., Hill, J., Befort, B.L., 2011. Global food demand and the sus-tainable intensification of agriculture. Proc. Natl. Acad. Sci. Unit. States Am. 108(50), 20260e20264.

Tittonell, P., 2014. Ecological intensification of agriculture - sustainable by nature.Current Opinion in Environmental Sustainability 8, 53e61.

Truong, T.T.A., Fry, J., Van Hoang, P., Ha, H.H., 2017. Comparative energy and eco-nomic analyses of conventional and System of Rice Intensification (SRI)methods of rice production in Thai Nguyen Province, Vietnam. Paddy WaterEnviron. 15 (4), 931e941.

Wan, N.-F., Cai, Y.-M., Shen, Y.-J., Ji, X.-Y., Wu, X.-W., Zheng, X.-R., Cheng, W., Li, J.,Jiang, Y.-P., Chen, X., et al., 2018. Increasing plant diversity with border cropsreduces insecticide use and increases crop yield in urban agriculture. eLife 7,e35103.

Wan, N.-F., Chen, J., Ji, X.-Y., Chac�on-Labella, J., Zhang, H., Fan, N.-N., Jiang, J.-X., Li, B.,2019a. Co-culture of multiple aquatic species enhances vegetable production incoastal Shanghai. J. Clean. Prod. 241, 118419.

Wan, N.-F., Li, S.-X., Li, T., Cavalieri, A., Weiner, J., Zheng, X.-Q., Ji, X.-Y., Zhang, J.-Q.,Zhang, H.-L., Zhang, H., Bai, N.-L., Chen, Y.-J., Zhang, H.-Y., Tao, X.-B., Zhang, H.-L.,Lv, W.-G., Jiang, J.-X., Li, B., 2019b. Ecological intensification of rice productionthrough rice-fish co-culture. J. Clean. Prod. 234, 1002e1012.

Wan, N.-F., Su, H., Cavalieri, A., Brack, B., Wang, J.-Y., Weiner, J., Fan, N.-N., Ji, X.-Y.,Jiang, J.-X., 2020. Multispecies co-culture promotes ecological intensification ofvegetable production. J. Clean. Prod., 120851

Woods, J., Williams, A., Hughes, J.K., Black, M., Murphy, R., 2010. Energy and thefood system. Philos. Trans. R. Soc. Lond. B Biol. Sci. 365 (1554), 2991e3006.

Yuan, S., Peng, S., 2017. Input-output energy analysis of rice production in differentcrop management practices in Central China. Energy 141, 1124e1132.

Zhao, L., Wu, L., Li, Y., Lu, X., Zhu, D., Uphoff, N., 2009. Influence of the system of riceintensification on rice yield and nitrogen and water use efficiency with differentN application rates. Exp. Agric. 45 (3), 275e286.